Monetary and Macro-Prudential Policies

Transcription

1 Monetary and Macro-Prudential Policies Jorge Roldos IMF-CEMLA Course Central Bank of Brazil, Brasilia October 213 This training material is the property of the International Monetary Fund (IMF) and is intended for use in IMF Institute courses. Any reuse requires the permission of the IMF Institute.

2 There is a danger that the macroeconomic models now in use in central banks have been constructed in the past as part of the war against inflation. The central banks are prepared to fight the last war. But are they prepared to fight the new one against financial upheavals and recession? The macroeconomic models they have today certainly do not provide them with the right tools to be successful. Paul de Grauwe, 28

3 Contents Monetary Policy in Inflation Targeting Regimes Case Study 1: Inflation Targeting and Financial Instability Monetary Models and Financial Intermediation Monetary and Macro-Prudential Policies Monetary versus Macro-Prudential Rules Assessing the Macro Impact of Basel III

4 I. Monetary Policy and Inflation Targeting Central banks have three main objectives: Price Stability Output/ Employment Stability Financial Stability Monetary Policy J.Roldos 4

5 I. Monetary Policy and Inflation Targeting Financial Stability has always been an objective for central banks Lender of last resort (LOLR) function Main reason for their creation in many cases After WWII, output and price stability became the main objectives In the 199 s, monetary dominance overtakes fiscal dominance in macro policies

6 Convergence in Macroeconomics Since the mid-198 s we saw Great Moderation A new macroeconomic synthesis emerged Macro analysis should use models with consistent intertemporal general equilibrium foundation (DSGE) Econometrically validated/calibrated Endogenous expectations Monetary policy should use interest rate to manage aggregate demand New Keynesian model used by many central banks

7 The New Keynesian DSGE Model Aggregate demand, IS equation x = Ex [ r Eπ r ], where x = y y * * t t t+ 1 t t t+ 1 t t t t is the output gap Expectations-augmented Phillips curve πt = βetπt+ 1 + κxt, where πt is inflation Policy (Taylor) rule rt = αrt 1 + (1 α)[ φπ π t + φxxt], where rt is the interest rate

8 The New Keynesian DSGE Model It is basically the IS-LM model with a Monetary Policy (MP) rule instead of LM (MM equilibrium) Macroeconomic equilibrium shown in next charts (for given inflation expectations) This three-equation NK model (Clarida-Gertler- Gali) became the backbone of central banks inflation targeting (IT) regimes

9 The NK-DSGE is an IS-MP Model

10 Case Study 1: Inflation Targeting May Lead to Financial Instability Based on Christiano et al. (27) There is empirical evidence that episodes of low goods price inflation are associated with high asset price inflation (stock price booms) Boom may reflect inefficiently loose monetary policy in an IT regime, when optimism about the future requires high real interest rates Problem of misusing the NK model

11 Periods of low inflation coincide with stock price and credit booms (U.S )

12 Periods of low inflation coincide with stock price and credit booms (U.S )

13 Summary Statistical Evidence

14 Case Study 1: Inflation Targeting May Lead to Financial Instability Boom-bust can be rationalized with NK model Problem is that model is a gap model: all variables as deviations from steady state And steady-state is not constant, should reflect the evolution of natural output, rates Moreover, news or signals of better future times could be an important driver of natural GDP, interest rates

15 The missing part of the 3-equation NK-model Natural or equilibrium output responds to productivity ( supply ) and leisure/labor ( demand ) shocks: * 1 yt = at τ t, where at is the productivity shock ϕ r = E ( y y ) = E ( a a ), and a evolves as: * * * t t t+ 1 t t t+ 1 t t a = ρa + u ; u = ξ + ξ ; ξ is "news shock" 1 1 t t 1 t t t t 1 t 1 The news shock is signal of future productivity

16 The missing part of the 3-equation NK-model Replacing evolution of technology into equilibrium interest rate: r = ( 1) a + * ρ ξ 1 t t t small large News of future technology improvement requires higher current real interest rate

17 The missing part of the 3-equation NK-model But in the NK model higher future productivity means lower future marginal costs and inflation mc π π r e e t+ j t+ j t t (Taylor Rule) and Taylor rule leads central bank to lower current nominal interest rate, inflating asset prices and creating a boom

18 Case Study 1: Inflation Targeting May Lead to Financial Instability Bottom line: ignoring the underlying structural economy, and/or relevant shocks, could lead to mistakes in model use CIMR (27) show that including a response to credit growth in the monetary policy rule improves welfare by avoiding boom-bust Why? : Credit growth responds to signal of higher future productivity, push rates higher Go back to themes of L-3, credit boom-bust

19 II. Monetary Policy Models and Financial Intermediation (FI) Early attempts to add FI to macro models: Two interest rates Rationale for FI IS curve flatter (FA) Shocks to credit supply important for business cycles (shifts to IS curve)

20 A Disruption to Credit Supply

21

22 II. Monetary Policy Models and Financial Intermediation Main challenge to NK DSGE (or any) model is that FI requires two types of agents: borrower (impatient) and lender (patient) Another challenge is the behavior of the FI, and the resource cost for the economy And a final one is keeping track of the dynamics of borrowing (credit) Some classes of models that have made progress on these are reviewed in what follows

23 Curdia-Woodford (29) Simple NK model with two additions 1. Heterogeneous households (borrower&lender) 2. Costly FI: resource cost and NPL Two factors affect the three equations: λ λ =Ω = s + δeω, where Ω is mg utility gap b s t t t t t t+ 1 t r r = s b =Ξ b + χ b b s ' ' t t t( t) t( ) t( ), spread = mg cost of lending, where Ξ ( b) is resource cost of lending and χ ( b) are NPLs t t

24 Curdia-Woodford (29) A fourth equation would have to be added: s bt = δ[1 + αsst + αbξ t] bt 1(1 + rt 1) / πt αbξ t( bt) + other b s (1 + r ) b = (1 + r ) d, is FI balance sheet t t t t cost of borrow The three-equation model suffices only in special case where credit spread evolves exogenously and FI uses no resources (spread is just a mark-up)

25 Curdia-Woodford (29) Even under such strict conditions, response to financial shock is equivalent to simultaneous shocks to: Natural interest rate Cost-push inflation Monetary policy shock Result: Monetary Policy Rule that incorporates credit spread is better (more on this below), but weight depends on persistence of the financial shock

26 Christiano, Motto and Rostagno (29) Model combines NK DSGE with capital accumulation, and financial accelerator (FA) Savings done by household, and risk neutral entrepreneur borrows to invest Need to add three new equations: 1. Optimal lending contract (menu) 2. Bank s zero profit condition 3. Law of motion for entrepreneur net worth

27 Christiano, Motto and Rostagno (29) Optimal contract where marginal revenue equals marginal cost of internal funds times external finance premium (EFP): αr QK = ρω ( ) R Q, where ρω ( ) is the EFP and R k it, + 1 k α 1 t+ 1 t t+ 1 t+ 1 t ω = [ MPK + (1 δ) Q ] it, + 1 t+ 1 t+ 1 Q t is the return on capital Individual, idioscincratic shocks averaged out

28 Christiano, Motto and Rostagno (29) Bank monitors borrower at default such that expected return equals cost of funds: [ ( ) ( )] k α Γ ω µ Gω Rt Qt 1Kt = RB t t, where µ is monitoring cost expected proj. success Combining both expressions we get credit spread as a function of leverage: R R QK = S = N k t+ 1 t t+ 1 t+ 1 t+ 1 SL [ ( ω)]

29 Christiano, Motto and Rostagno (29) The evolution of entrepreneur net worth is: N = R Q K RB = k t+ 1 t t 1 t t t N = ( R Q R ) K + R N k t 1 t t+ 1 t t t t Nt Asset Returns-Debt Payout Entrepreneur s net worth grows with leverage Incentive to max leverage, mitigated by increasing spread

30 Christiano, Motto and Rostagno (29) Log-linearizing the two equations we get: fi fi s = χ( q + k n ) + ε, where ε is shock to FI t t t t t t n = Lr ( L 1)( s + r π ) + θn + ε, k nw t t 1 t 1 t 1 t t 1 t where ε nw t is shock to net worth The financial shock could also be an increase in the volatility of the firm s profitability A (tail) risk shock, on real sector This type of shock sacounts for a large share of business cycles in US and EU

31

32

33 Christiano, Motto and Rostagno (29) First shock to entrepreneurial risk leads to: Increase in spreads from 22 to 24 bps Reduction in D/N from.92 to.8, 13% fall in loans Second shock, Δ in monitoring/bankruptcy cost: Spread or EFP falls 64 bps (demand dominates) Similar reduction in D/N or lending Monetary Policy tightening of 25 percent: Spread rises by 34 bps Lending falls 1 percent

34 Christiano, Motto and Rostagno (29) Model with micro-founded FI shows how moves in spread may be small compared to those in lending (if dem&supply move same direction) Models with spread only (or without explicit contract for lending could be misleading) Again, adding credit growth to Monetary Policy rule improves welfare One draw back of CMR: intermediary has no capital; no bank failure or runs

35 Gilchrist and Zakrjsek (211) Similar NK with FA, but shock to FI (not real project) χ QK t t+ 1 exp( σ t), and the shock to FI follows: Nt St = σ = (1 ρ) σ + ρσ + ε t t 1 t An Δσ increases cost of FI for given leverage Model matches the recent U.S. crisis, and spread-augmented MP rule stabilizes GDP (with a bit of Δ inflation)

36 Gilchrist and Zakrjsek (211)

37 Gilchrist and Zakrjsek (211)

38 Gilchrist and Zakrjsek (211)

39 II. Monetary Policy Models and Financial Intermediation (FI) cont. Taking stock: NK DSGE + FA provides a good degree of amplification Dynamics of non-financial sector balance sheet and leverage micro-founded; spreads are counter-cyclical, good data fit But no meaningful FI, same with collateral constraints framework (see next table) Next class of models explicitly introduce a costly banking sector

40 Source: IMF/WP/

41 Gerali et al. (21) A NK DSGE model with an explicit banking sector and collateral constraints Emphasis on supply-side of credit markets Entrepreneurs and impatient households (HHs) borrow, patient HHs lend deposits to banks Banks are monopolistically competitive, setting sticky loan and deposit rates; bank capital One wholesale branch, two retail branches

42 Gerali et al. (21) Motivation: Interest rate spreads affected by degree of competition and bank capital costs

43 Gerali et al. (21) Entrepreneurs and impatient HHs borrow subject to collateral (LTV) constraints: Rb( i) me[ q k( i) π ], where m is the LTV b k t t t t t+ 1 t t+ 1 t Banks lending and deposit rates are set taking into account future values of the policy rate R > R > R b d t t t 5.3>3.6>2.4 (percent) ; lending, policy and deposit rates

44 Gerali et al. (21) Bank capital is accumulated out of retained earnings π K = (1 δ ) K + j ; where j are retained earnings b b b b b t t t 1 t t Banks max profits, and is costly to adjust banks capital ratio (inverse to leverage L) 2 b b d κ K b t b b max RB t t RD t t υ Kt ; υ is targeted CAR 2 B t

45 Gerali et al. (21) Wholesale bank lending rates then depend on: R 2 b b b d Kt b K t t = Rt κ υ Bt Bt Bank spread depends on bank leverage in contrast to CMR (on firm s leverage) 2 b b w b d Kt b K t t = t t = κ υ Bt Bt S R R,

46 Gerali et al. (21) Macroeconomic shocks affect banks profits and capital, with feedback to the real economy Banks affect monetary transmission, attenuating effects on real economy (sticky interest rates) Banks also are subject to financial shocks (spreads, collateral and capital) that affect the economy Financial shocks explain large share of GDP fluctuations in (next)

47

48 Gerali et al. (21) Model suited to study effects of bank capital loss (crisis) as well as recapitalization efforts After capital loss, banks attempt to rebuild balance sheet: cut lending and increase spread Firms reduce investment, but increase capacity utilization and labor demand Inflation lead central bank to tighten marginally Stress scenario with added ΔCAR (regulation)

49 Gerali et al.: Loss of Bank Capital

50 Gerali et al.: Loss of Bank Capital

51 III. Monetary and Macro-Prudential Policies The models sketched before allow for a number of applications In particular, analysis of monetary and macroprudential policies Next, three applications Monetary versus Macro-Prudential Rules Impact of Basel III on GDP Counter-cyclical capital adequacy ratios

52 Monetary versus Macro-Prudential Rules Based on Kannan, Rabanal and Scott (29) Key question: should central bank willing to mitigate boom-bust cycles move policy rate in response to credit or asset prices/spreads? Answer is YES, but for purely financial shock the macro-prudential instrument has comparative advantage over MP rate Discretion needed since is hard to know the source of shocks

53 Kannan, Rabanal and Scott (29) Model has non-durable and durable good (housing), only real asset Patient (lender) and impatient (borrower) HHs FI set lending rate as function of LTV, s.t. financial shock and macro-prudential instrument: b b h R = ν RF( B / QK ) τ ; where F (.) >, F (.) > ; t t t t t t t ν, τ, are financial shock and macroprudential instrument t t

54 Kannan, Rabanal and Scott (29) Compare four policy regimes: Taylor Rule Augmented Taylor Rule (with credit growth) Augmented Taylor Rule + Macro-prudential Optimization [Augm.TR + Macro-prudential] The augmented TR and macro-prudential are: Rt = γrrt 1+ (1 γr)[ γπ π t 1+ γ yxt 1+ γbbt 1], augm. TR τ = τ( B ), macroprudential instrument t t 1

55

56 Effect of a Financial Shock

57

58 Effect of a Productivity Shock

59 Kannan, Rabanal and Scott (29) Bottom line: comparative advantage dictates that macro-prudential (M Policy) instrument should be used against financial (technology) shock Critical to be able to isolate shocks The higher the incidence of financial shocks in an economy, the more important the relative use of the macro-prudential instrument (LTV, RR)

60

61 Case Study 2: Shocks to World Interest Rates and Capital Flows In the aftermath of Lehman Bros. collapse, most countries cut interest rates sharply Same in major EM countries, but some of them cut Reserve Requirements even more (and earlier); especially Brazil and Peru (chart) What is the best response to this shock (and associated fluctuations in capital flows)? What if rates go back up (Fed normalization )?

62 MP Rate and Reserve Requirements

63 Shock to the World Interest Rate

64 Case Study 2: Shocks to Global Interest Rates and Capital Flows In Christiano et al, future (positive) productivity shocks increase the natural rate of interest But their was closed-economy model: for shock to global rates we want an open-economy model A few available, focus on other shocks or on response with capital controls An exception is Medina and Roldos (forthcoming IMF/WP), next

65 Overview of the Model Small open economy Two differentiated tradable goods: Home and Foreign Nominal friction: price rigidity a-la-calvo (1983) Financial friction: financial accelerator plus a firesales amplification mechanism Solvency: loans default (Bernanke, Gertler and Gilchrist, BGG, 1999) Liquidity: real and financial resources are needed to liquidate distressed assets (extension of Choi-Cook, 212)

66 Firms Labor Market Capital Producers Entrepreneurs l R K R Households D R Liquidity Intermediaries IB R Lending Intermediaries K l IB D R > R > R > R

67 Figure 2 Timing of events Period t Period t+1 Using net worth and loans Entrepreneur buys new eop capital K t + 1 from capital goods producers B ω + 1 t After realization of shock t Entrepreneur supplies capital services Entrepreneur sell undepreciated (1 δ ) Kt+ 1 to capital producers, repays loan to lend-intermediary Production and Consumption Capital producers buy new, undepreciated and restructured capital Liquid-intermediary takes deposits, lends to interbank market and CB; Provides liquidity services Lend-intermediary sells defaulted capital to liquidintermediary; lends to finance next period capital

68 Financial Intermediaries and Spreads The two financial intermediaries summarize the main functions of a financial system: Provision of credit services Provision of liquidity services Both functions contribute to a wedge between deposit and lending rates: Endogenous spreads will interact with monetary policy rate and macro-prudential instrument

69 Credit Intermediary (1) Provides loans to entrepreneurs under a BGG contract that solves the agency problem The average return to capital for the entrepreneurs is capital is: R = VMPK + (1 δ ) Q ; where VMPK is the rental rate K t+ 1 t+ 1 t+ 1 t+ 1 Qt and Q t the price of capital

70 Credit Intermediary (2) Threshold condition for lending rate: R ϖ K Q = R B, where R is loan rate and k l l t+ 1 t+ 1 t+ 1 t t+ 1 t t+ 1 is the loan Zero profit condition for credit intermediary: B t ϖ t+ 1 IB ( VMPK ) [1 Φ ( ϖφ( )]; R B) +, + (1 δ) FS k ωd ω σ ω = R B l t+ 1 t+ 1 t t+ 1 t+ 1 t+ 1 t+ 1 t Revenue if loan repaid Revenue if default Cost of Funds where Φ( ϖ ) is the probability of default and FS t + 1 t+ 1 is the fire-sale price of the defaulted capital

71 Credit Intermediary (3) Define: µ t+ 1 = ( Qt+ 1 FSt+ 1)(1 δ ), VMPK + (1 δ ) Q t+ 1 t+ 1 Then, zero profit condition for credit intermediary is: l k IB [1 Φ Φ( ( ϖ t+ 1; )] Rt+ ) 1Bt + ( 1 µ t+ 1) Rt+ 1Qk t t+ 1 ωd ω σ ω = R t + 1Bt, Revenue if loan repaid Revenue if default Cost of Funds In contrast to BGG, cost of default is endogenous (and countercyclical); in recession: Prob (Default) increases Recovery rates fall (cost of default increases) ϖ

72 Liquidity Intermediary (1) Demand for liquidity services: lqt =υkdt,, where lqt are liquidity services To provide liquidity services, LI requires real and financial ( excess reserves ) resources 1 αlq 1 α t = t t xr lq min[ n ; xr ], where n is in units of final goods, and xr is "excess reserves"

73 Liquidity Intermediary (2) Maximize benefits derived from the use of deposits: IB MA RE D Dt max (1 s ) R + s R R Pn D,,( / ) R t t t t t t t ns DP t Allocation of funds: MA s t s t 1 Reserve Requirement Exc. Reserves Lending to Interbank Market

74 Liquidity Intermediary (3) Optimal condition for liquidity intermediary: (1 α ) lq lq IB t R = g R xr D t t t t IB MA 1 D MA t = (1 t ) [ t t ] R s R s Fire sale (or cash-in-the-market ) price: ( ) (1 )(1 ) [ ] FSt = ηkqt υ ft + gt + ηk δ Et sdt+ FSt+ Equilibrium in the interbank market: B = (1 s ) D t t t 1 1

75 Nominal friction Wholesale producers sell differentiated goods, a composite of domestic and imported goods Set sticky prices a-la-calvo Phillips curve: β χ log(1 + π ) = E [log(1 + π )] + log(1 + π ) p t t t+ 1 t 1 1+ βχ p 1+ βχ p marginal cost is: mgcr (1 ) (1 φp)(1 βφp) mgcr t + log, where φp (1 + βχ p ) mgcss 1 θ d * Pyt, ep t t t = αd + αd Pt Pt 1 1 θ d 1 θ d

76 Aggregate Equilibrium Aggregate (domestic) demand is given by da = c + c + inv + n t t kt, t t And is equal to the supply of final goods (adjusted by the price distortion or nominal friction disp ) da = ( disp ) y 1 t t st,

77 Summary of Spread Determinants The spread between the lending rate and the deposit rate is determined by leverage and liquidity conditions; policy instruments: l d MA R = R f QK / N, µ Q FS ; s macro leverage liquidity pru policy instr.

78 Calibration Macro-parameters: standard Financial system parameters: (i) annual default rate of 3 percent (in line with BGG); (ii) a leverage ratio of 4 percent (mid-point between BGG and estimate from Gonzalez-Miranda (212) ); (iii) an average cost of liquidation of 6 percent These parameter values imply: R K = 13.7% > R l = 6.6%, R IB = 4.5%, R D = 4% (in annual basis) a recovery rate of around 36 percent entrepreneurs debt/credit is 55 percent of annual GDP deposits, as percentage of annual GDP, is 61 percent Excess reserves as percentage of annual GDP of.15 percent

79 Alternative Policies We consider four alternative regimes 1. Standard Taylor rule and constant reserve requirement (RR) 2. Inflation Targeting and constant RR 3. Augmented Taylor rule and constant RR IB R t log = ψπ log(1 + πt) + ψ ylog( yt) + ψblog( bt) 1+ r 4. Inflation Targeting and countercyclical RR MA s t xr t log φxr log MA = s xr

80 Natural Price rigidities, financial frictions+it regime Price rigidities, financial frictions+augmented Taylor rule forces and more debt/credit 1.5 GDP 8 Investment 2 Real exchange rate 1 interbank rate Dev. from SS aggregate demand 4 default rate inflation rate 4 Tobin Q Dev. from SS fire sale price 4 deposits (% SS GDP) 4 ent. debt (% SS GDP) 1 deposit rate Dev. from SS excess of reserves (% SS GDP) 1 Foreign debt (% SS GDP) 3 networth (% SS GDP) 2 Loan rate Dev. from SS Quarters Quarters Quarters Quarters

81 Alternative policies (w/o MaPP) 1 Reserve requirement (% Deposits) 2 Cost of Liquidation 3 Recovery Rate Dev. from SS Quarters Quarters Quarters Policy Framework Welfare Losses (+)/ gains (-) expressed relative to the deterministic SS welfare in terms of SS consumption Standard Taylor rule % IT regime % Augmented Taylor rule %

82 Natural Price rigidities, financial frictions+it regime Price rigidities, financial frictions+augmented Taylor rule Price rigidities, financial frictions+it regime & countercyclical res. req. 1.5 GDP 8 Investment 2 Real exchange rate 1 interbank rate Dev. from SS aggregate demand 4 default rate inflation rate 4 Tobin Q Dev. from SS fire sale price 4 deposits (% SS GDP) 4 ent. debt (% SS GDP) 1 deposit rate Dev. from SS excess of reserves (% SS GDP) 1 Foreign debt (% SS GDP) 3 networth (% SS GDP) 2 Loan rate Dev. from SS Quarters Quarters Quarters Quarters

83 Higher Welfare with Countercyclical RR Policy Framework Welfare Losses (+)/ gains (-) expressed relative to the deterministic SS welfare in terms of SS consumption 3 Augmented Taylor type rule % 4 IT regime and Countercyclical RR % 3 Reserve requirement (% Deposits) 2 Cost of Liquidation 3 Recovery Rate Dev. from SS Quarters Quarters Quarters

84 MP rate: coordinate with Macroprudential tool Tinbergen, Mundell principles MP rate has to follow the natural rate and be adjusted by the RR 1 interbank rate

85 Robustness: Dollarization (1) Include foreign debt (denominated in dollars) for entrepreneurs and credit intermediaries More financial volatility, but same policy ranking Policy Framework Welfare Losses (+)/ gains (-) expressed relative to the deterministic SS welfare in terms of SS consumption 1 Standard Taylor rule % 2 IT Regime % 3 Augmented Taylor rule % 4 IT regime and Countercyclical RR %

86 Robustness: Dollarization (2) Loans to entrepreneurs are denominated in dollars. This also rises the volatility Taylor rules dominate IT, MaPP improves the welfare even more. Policy Framework Welfare Losses (+)/ gains (-) expressed relative to the deterministic SS welfare in terms of SS consumption 1 Standard Taylor type rule % 2 IT Regime % 3 Augmented Taylor type rule % 4 IT regime and Countercyclical RR %

87 Dollarization (2) Natural Price rigidities, financial frictions+it regime Price rigidities, financial frictions+augmented Taylor rule Price rigidities, financial frictions+it regime & countercyclical res. req. 1.5 GDP 1 Investment 2 Real exchange rate 1 interbank rate Dev. from SS aggregate demand 1 default rate inflation rate 1 Tobin Q Dev. from SS fire sale price 6 deposits (% SS GDP) 6 ent. debt (% SS GDP) 1 deposit rate Dev. from SS excess of reserves (% SS GDP) 1 Foreign debt (% SS GDP) 4 networth (% SS GDP) 1 Loan rate Dev. from SS Quarters Quarters Quarters Quarters

88 Robustness: Wage rigidities Only a fraction of workers adjust their wages to labor market condition every period (Blanchard- Galí, 27) This increases the responses of asset prices and defaults. Distortions reinforce each other and MaPP delivers more welfare benefits Losses (+)/ gains (-) expressed Policy Framework Welfare relative to the deterministic SS welfare in terms of SS consumption 1 Standard Taylor type rule % 2 IT Regime % 3 Augmented Taylor type rule % 4 IT regime and Countercyclical RR %

89 Case Study 2: Conclusion A countercyclical macroprudential policy is better suited to manage the volatility of world interest rates and associated capital flows Conclusion is robust to different combinations of nominal and financial frictions (other MaPP instruments? Probably, see next) Inflation Targeting continues to be the main monetary policy objective, but MP rate must accommodate moves in the natural interest rate and reserve requirements

90 Bianchi: Equivalence Between Reserve and Capital Adequacy Requirements

91 Bianchi s Equivalence Result

92 Macroeconomic Costs of Higher CAR Based on Rogers and Vlcek (IMF WP/11/13) Model similar to Gerali et al The macroeconomic impact of a 2% increase in CAR depends on: Banks optimal response: increase spreads, reduce dividends or lending Monetary policy response Implementation period

93 Macroeconomic Costs of Higher CAR The increase in CAR happens gradually over two years The least costly option is to reduce distribution of bank dividends But is not enough: next comes an increase in spreads Finally, last resort is a reduction in lending (with maximum negative impact on GDP)

94

95

96

97 Macroeconomic Costs of Higher Two effects: Liquidity Requirements Lower bank revenues, thus higher spreads Lower risky assets, thus less capital needed The last effect means that liquidity and CAR become complements; interaction effects are very important For a 25% increase in liquid assets, spreads increase only about 5 bps

98

99 BIS Assessment of Δ CAR Conducted with a battery of models, across several countries Results are similar to those just shown with IMF model Typically, output reduction is rather small Contrast with industry estimates

100

101 Counter-cyclical Capital Buffers Using models similar to above examples, IMF (GFSR September 211) estimates role of counter-cyclical capital buffers on financial and macro-economic stability Baseline is an asset price bubble that builds up financial imbalances Volatility of GDP and trade balance are much lower with counter-cyclical buffer (next)

102

103 Credit Boom in Open Economy Model with simple FA (Unsal, IMF WP/11/189) where an improvement in external conditions leads to capital inflow Imposing regulatory premium complements monetary policy (theme of afternoon workshop)

104 Final Thoughts on Model Use Models reviewed (and others) are showing promising results for a number of macroprudential policy analyses The sources of shocks hitting the economy are critical Models could be used to deliver implications for other variables, to help decide whether an increase in credit growth is an imbalance (say, bubble) or genuine (due to productivity)

105

106 Final Thoughts on Model Use Successful use of models require deep understanding of transmission mechanisms, relation to data, and above all judgment However, some models ignore a number of important issues, many raised by recent crisis In particular, increased risk-taking, bank runs and asset price bubbles Need to be complemented with indicators of build-up of imbalances and financial intelligence

107 Final Thoughts Monetary and Macro-Prudential Policies are broadly complementary Models are useful, if used with judgment Financial Stability is a central bank objective, to be achieved with appropriate weight to monetary and macro-prudential tools It is critical to identify the source of shocks: financial or others; models and indicators useful for this as well

108 Thank You!